Earlier this month, I attended Forrester’s inaugural Data Strategy and Insights Conference. During the event, I had the opportunity to present with Piyush Jain, Director of Data Governance at Progressive Insurance, in a session entitled, “Transforming the Data Culture: Driving Data Value with Collibra at Progressive Insurance.” Much of what was covered during our session were themes that were discussed throughout the event including how trustworthy, governed data can increase usage throughout the organization and accelerate AI implementation.
I left the conference with three key takeaways:
1. AI increases the importance of data governance. Most of the attendees I spoke with had an AI project running in pilot or in production for a narrow segment. But the plans to scale up AI spending were not there yet due to a lack of trust in AI. Governance was often discussed as a potential solution, and understanding how to align data governance and AI governance is a key component. Having complete confidence in data is necessary to have confidence in automated decision making.
Key takeaway: Data governance efforts must include AI. If you haven’t yet started a governance program, then now is the time. Without one, there will be no trust in AI. With a governance program, you will also get the benefits of data visibility and quality. If you have a governance program, be sure to include data relevant to AI projects. Training data, input sources, and the output of the process all must be governed data whose quality is monitored.
2. It’s all about the data fabric. Much of the discussion on the second day centered around the development of a data “fabric.” This fabric, as defined by Forrester, is the collection of data capabilities that organizations need to have easy access to their data including a data catalog and business glossary. These two work together with the glossary enabling users the ability to find data that they need according to its meaning. The data catalog then links this meaning with the data for provisioning. The catalog also exposes the content of the data, like its usage, lineage, and the opinions of other users.
Key takeaway: Organizations must make their data visible and have defined processes for provisioning data. Visibility into data is based on knowing its meaning. A business glossary linked to technical metadata is the best way to capture this meaning. This is the cornerstone of an organization’s data fabric and it must be independent of the use or storage of the data. This will make it possible to find the data and use it anywhere that it will add value to the organization.
3. Everyone is doing it. We met many people at the event and it was clear that using data as part of a digital transformation was occurring in each person’s day-to-day life. The questions focused on execution tactics and how to make data programs a reality. This focus is a reflection of maturing data practice. Leading organizations are innovating quickly and their less-aggressive peers are now implementing those best practices. The three data practices, visibility, integrity, and quality, are the foundation of digital initiatives. Only when you can see all the data you have can you determine the most profitable uses for it. When you are sure that you can use the data safely, you can make intelligent decisions about how customers will react to your brand. And when the quality of the data is well known, you can be confident that it is fit for purpose.
Key takeaway: Data initiatives will speed up as established best practices increase their efficiency.
Data is making its way into nearly every aspect of the business and is the foundation of digital transformation. Your catalog and governance efforts should be tightly linked to your digital efforts. This will help you across all three practices: visibility, integrity, and quality.